The Victory Lab (25 page)

In November 2003, Podhorzer began three-week cycles, during which he would send out three different mail pieces and compare their effectiveness. Not only did the method allow the AFL to save its resources for the most effective messages; the large-sample polls also gave Podhorzer new insight into those who were really susceptible to having their minds changed. The polls showed him with increased precision what could make voters move away from Bush, and also groups that were resistant to arguments; with the latter, he went back and tested different mailers to see if other language might resonate more strongly. Overall, only between one-third and one-fifth of people in Podhorzer’s target universe moved in the way he had expected they would. Many did not show any effect, and some seemed turned off by the mailers. “It was obvious,” says Podhorzer, “that we were wasting a lot of our resources communicating with people who would never support us or people who would support us even if we never talked to them.”

As the experiment got under way, Podhorzer became increasingly frustrated with how little meaningful information he had about the voters he was trying to reach. The AFL’s conventional polling recorded what voters thought of the candidates and what issues were most important to them, but it revealed almost nothing about what information could change their minds. “You had someone answer a survey and say global warming is their most important issue, but sending them a piece saying Al Gore is for fighting global warming is useless,” says Podhorzer. “Unless their head’s under a rock they know Al Gore is against global warming.”

When a pollster asked if someone would be more or less likely to vote
for a candidate in favor of shipping jobs overseas—a typical way of auditioning what was then a promising line of attack against Bush—they would often hear from voters across the board that it made them “less likely.” But when the AFL sent out a draft leaflet about Bush’s free-trade policies, it turned out to have little impact on the autoworkers who received it. The knowledge of factory job loss was “baked in” to their impressions of Bush, as Podhorzer liked to put it: the workers already knew what the union wanted them to think about Republican trade policy. They liked or disliked Bush regardless. But other groups, like construction workers and Republicans, did not know as much. A piece of mail that gave them information turned out to be persuasive in changing their attitudes toward Bush. Experiments allowed Podhorzer to see which voters actually moved, not just those who said they might.

BUSH’S REELECTION
, and the fear that the newfound Republican vote-hunting mettle behind it might presage a generation out of power for Democrats, brought new urgency to the left’s previously fitful efforts toward innovation. On a cold, rainy afternoon in late November 2004, Laura Quinn called Debra DeShong and asked her to come to the consulting firm Quinn owned. Quinn said she had a Christmas gift for her former Democratic National Committee colleague, but it became quickly evident to DeShong that a wrapped iPod was not the only thing Quinn was eager to share. Quinn’s desk was covered with newspaper and magazine clips about how Bush had won, many of them lionizing Karl Rove, whom Bush had described the morning after his victory as “the Architect.”

It is common for operatives to spend much of their time trying to figure out what the opposition is up to, but by the end of 2004 Quinn’s fixation on Rove had risen to the level of obsession. She kept easily accessible on her computer desktop the video of
a 1972 CBS News report in which Dan Rather visits the Committee for the Re-Election of the President to
marvel at the latest techniques being used to support Richard Nixon’s reelection, including the earliest efforts at data-driven direct-mail fund-raising. “If you’ve ever contributed to the Republican Party or subscribed to a conservative magazine, or purchased Idaho Steaks through the mail or written away for a dry washcloth for your car, chances are your name is on the computer,” Rather intones. “And if you haven’t been asked to contribute to the Nixon campaign, you will.” The correspondent then travels to the basement, where he meets the executive director of the College Republicans, busy plotting Nixon’s youth registration efforts. “Young people have got to reach other young people, and that’s what we’re seeking to do,” says a twenty-one-year-old Karl Rove, decorated with glasses, a tie and vest combination, and luxurious sideburns. “It’s just so fun,” Quinn would say right before pressing play, a rare turn toward girlishness from a woman who regularly impressed colleagues and intimidated rivals with her intense, and occasionally joyless, focus.

Now, while others who had played a role in Kerry’s campaign were scattered on tropical beaches trying to put 2004 behind them, Quinn found a sense of purpose in her pile of Rove clips. She marveled at the way he had outfoxed the left at the aspects of campaigning where it had claimed mastery, and she was intent on reverse-engineering his methods from the general descriptions that had appeared in newspapers. Upon arriving at Quinn’s office, DeShong was surprised to find her friend unusually chipper. “She was very upset about losing the election, as we all were,” DeShong recalls. “But she was so excited that she had figured out what Rove had done.”

Looking for “the mark of Rove,” as some took to calling it, had become a popular pastime for many Democrats.
In June 2002, a Senate staffer crossing Lafayette Square, just across Pennsylvania Avenue from the White House, came across a CD-ROM. Upon examination, the mysterious find revealed unusually compelling contents: two PowerPoint presentations that Rove and White House political director Ken Mehlman had delivered to a Republican gathering at the nearby Hay-Adams hotel. The slides, which contained a forecast of that fall’s midterm elections, were
given to a Capitol Hill columnist and made news because they revealed a White House more pessimistic about the chances of two Republican Senate candidates than officials let on publicly. But the actual presentations gave away little except suspicions that the accidental loss of the disk was not what it appeared. (During the 2000 campaign, a videotape of Bush’s debate-prep sessions had arrived by mail at Gore’s headquarters, and allegations of older false-flag operations had followed Rove from Texas to Washington.) Could the famously disciplined Bush operation really be so sloppy? Or were they hoping it would be found?

That winter, another set of PowerPoint slides found its way into Democratic hands. The computer file lacked the concise founding narrative of the Lafayette Square disk, and even the most fevered conspiracist couldn’t imagine why Rove would want to see it public. The presentation had been designed by Blaise Hazelwood a year earlier, before the RNC’s 2002 winter meeting, with a complete set of findings from the party’s 72-Hour experiments. The material alerted Democrats that the Republicans had turned their attention to turnout and had developed an intellectual infrastructure for their field operations that towered over anything the institutions of the left had ever built. “You would have thought this was nuclear code, or a DNA sequence had been cracked,” says Tracy Sefl, the DNC’s deputy research director.

Hardly any details about Alexander Gage’s microtargeting project had made it into the press during the election. Instead, the stories that had come out about the implementation of 72-Hour tactics had been ones Republicans wanted told: about the “multilevel marketing” that was making Bush a ground-level force and helping to empower volunteers. Quinn and her allies suspected that Republicans had sharpened their approach to voter contact, but never knew how. They rigorously collected distinctive pieces of mail—the 2004 brochure sent to New Mexico Spanish speakers with a chart contrasting the positions of Bush and Kerry on abortion and gay marriage, over the slogan “Vote Sus Valores”—and speculated about the methods behind them. After the election, Rove and other advisers revealed what
they had been up to, taking what Democrats described as a “victory lap” for the so-called microtargeting methods that made them possible.

Even though no Democrats had used the word
microtargeting
, several party operatives had arrived separately at the same basic insight as Gage. They knew large-scale surveys could isolate the influence of personal characteristics that were combining in ways imperceptible to traditional polls, and how to track back to find specific individuals who fit those categories. In 2001, two former Gore advisers at Boston’s Dewey Square Group tried to market a clustering system called Fusion, modeled on products Claritas had developed for the commercial market. The next year, Malchow sold a CHAID program, similar to the one he had used for Ron Wyden in Oregon six years earlier, to Democratic-coordinated campaigns in three states.

But a small group of veterans of the Kerry operation known as the BullsEye felt the greatest frustration as they read the press accounts of Gage’s triumph. The BullsEye, which commandeered a small room at headquarters hidden off to the side of the communications war room, was designed to be the tactical hub of Kerry’s general election campaign, where a constant pulse of data from battleground states would help redirect the candidate’s plane or a communications blast to the areas that needed it most. One of the most important tools at their disposal was a master database that pollster Mark Mellman had built to contain all the interviews conducted by Kerry’s retinue of far-flung survey takers throughout the year. Instead of relying merely on a rolling series of state-by-state snapshots, the BullsEye could aggregate tens of thousands of respondents, track each of them back to a record in the voter file, and look for unusual patterns.

Before the Iowa caucuses, Ken Strasma, a former analyst for the National Committee for an Effective Congress, had conducted a ten-thousand-person poll to build a statistical model that could identify likely Kerry supporters for the candidate’s voter contact operation to target and turn out for the caucuses. In the general election, Mellman’s database would make it possible for Strasma to design similar models for voters nationwide. But there was little interest from the campaign’s leadership; Strasma
didn’t even speak with the campaign manager, Mary Beth Cahill, until well after Kerry’s dramatic Iowa victory. “It was just one of the things that wasn’t on her radar screen,” says Strasma.

Mellman’s database yielded important tidbits that helped to shape message strategy, such as the insight that voters who had contributed to disease-related charities were a promising audience for Kerry. As a result, the campaign gave Ron Reagan, the former president’s son, a prominent convention speaking slot to argue that Kerry’s support for stem-cell research could help cure the Alzheimer’s that had crippled Reagan’s father. But when it came to talking to voters directly, Kerry’s targeters had trouble getting state directors to put aside their precinct-based strategies and use the new individual-level profiles instead. At a weekend-long retreat held in Maryland for Kerry’s state-level operatives, only one hour was devoted to introducing and explaining Strasma’s new targeting program. “They would fight with me,” says Mellman. “The resistance was substantial, because people are used to looking at things in the way they’re used to looking at things.”

That resistance finally began to crack after the election, as victorious Republicans flaunted their fine-grained knowledge of the electorate. In one particularly potent example, Gage told how his party had turned even shopping patterns into political intelligence, discovering that bourbon drinkers leaned Republican while cognac sniffers were more likely Democrats. “It just scared the shit out of all the Democrats,” says Malchow. “The best way to get anyone to do anything on the Democratic side—and I’m sure it’s the reverse on the Republican side—is to tell people that the Republicans are doing it. It doesn’t matter: the Republicans could be doing something completely stupid, but if you tell the Democrats they get scared and think they should do it. They all think the Republicans are smarter than they are.”

Malchow, Strasma, and others started to look closely at some of the consumer variables that journalists loved to highlight in their stories, such as the idea that knowing what type of car someone drives will say something new about how they vote. Even Malchow got scared: “The Republicans
learned how to do car ownership!” When he looked more closely, though, he realized that even if Gage had managed to acquire the data and append it to a voter file, it couldn’t possibly be a usefully predictive variable for political contact. The car most overrepresented by Republican drivers was the Jaguar: 59 percent of Jag owners were Republicans. (Among Democrats, Subarus were strongest—43 percent of Subaru owners were Democrats.) But Jaguars amounted to only one-half of a percent of the total U.S. auto market, and there were plenty of other ways to identify older, ostentatiously wealthy suburbanites (like their addresses) than mining auto registration records for hints. A lot of the most colorful examples, Malchow concluded, were hype.

As Democrats learned more about the scale of what the Bush campaign had done, they realized that the opposition’s edge wasn’t about a particularly potent set of consumer files it had acquired but rather the political structure they had built around them. “On the Republican side, the RNC was so much stronger than the DNC was at the time,” says Hesla. No Democrat, and certainly not Kerry, had invested as much in individual-level targeting as Bush had, or did so early enough to integrate it fully into the campaign’s operations.
A
Washington Post
analysis of the $2.2 billion spent on the presidential campaign—split almost evenly between efforts on behalf of Bush and Kerry—concluded that Bush’s $3.25-million contract with Gage’s firm TargetPoint was among the best money spent that year. The
Post
story pointed to the increase in Bush’s turnout in Ohio, and included one quote from an anonymous Democratic operative declaring that the party’s targeting power was a full election cycle behind the Republicans’. “They came into Democratic areas with very specific targeted messages to take Democratic voters away from us,” Terry McAuliffe told the
Post
.